The concept of a router appears very simple, but indeed it’s quite remarkable. A router does not necessarily know, where a destination is actually located, but it may know at least the direction. When it receives a packet, it reads the address label and looks-up its table, whether it knows something about the neighborhood or larger area of that address. In most cases it will not know the exact location, but it often knows someone else, who could know more. And then it forwards the mail to the next hop so to speak in the hope, that this will be a step into the right direction.
Simple routers do just that, forwarding packets to their next neighbors based on statically configured routing tables. But the more advanced routers do more: they let others know, what they know, they talk. This dynamic routing allows a chain of routers to spread the news about a possible location, before any packet, searching for it, has even arrived. The result of this grapevine is a communication network, which is more than all its single members. There is more than one way, how various dynamic routing protocols realize this. But the mere principle implies, that the community of the routers mirror the whole network, if not the whole world, to some degree. They form an information network.
The concept of routing became extremely valuable in combination with the IP protocol, which is at the heart of nearly every larger network today and constitutes the underlying technology of the internet. Other than most earlier network protocols, IP does not depend on any central monitoring instance and its address scheme has no clear hierarchy or limit. The IP protocols was made for a open, flat, potentially endless network without any central brain or database. As no single instance or member of the network knows everything, routing is fundamental for finding the way.